geometry david a brannan matthew f esplen jeremy j gray

602
8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray http://slidepdf.com/reader/full/geometry-david-a-brannan-matthew-f-esplen-jeremy-j-gray 1/602  B  r  a  n  n  a  n  ,  E  s  p  l  e  n  ,  G  r  a  y  G  e  o  m  e  t  r   y   S  E  C  O  N  D   E  D  I  T  I  O  N This richly illustrated and clearly written undergraduate textbook captures the excitement and beauty of geometry. The approach is that of Klein in his Erlangen programme: a geometry is a space together with a set of transformations of the space. The authors explore various geometries: affine, projective, inversive, hyperbolic and elliptic. In each case they carefully explain the key results and discuss the relationships between the geometries.  New features in this Second Edition include concise end-of-chapter summaries to aid student revision, a list of Further Reading and a list of Special Symbols. The authors have also revised many of the end-of- chapter exercises to make them more challenging and to include some interesting new results. Full solutions to the 200 problems are included in the text, while complete solutions to all of the end-of-chapter exercises are available in a new Instructors’ Manual , which can be downloaded from www.cambridge.org/9781107647831 . Praise for the First Edition ‘To my mind, this is the best introductory book ever written on introductory university geometry… Not only are students introduced to a wide range of algebraic methods, but they will encounter a most pleasing combination of process and product.’  P. N. RUANE, MAA Focus ‘… an excellent and precisely written textbook that should be studied in depth by all would-be mathematicians.’  HANS SACHS,  American Mathematical Society ‘It conveys the beauty and excitement of the subject, avoiding the dryness of many geometry texts.’  J. I. HALL, Michigan State University Geometry DAVID A. BRANNA MATTHEW F. ESPL JEREMY J. GRAY SECOND EDITI

Upload: ketchumbud

Post on 07-Aug-2018

223 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    1/602

     B r  a n n a n , E  s  p l   e n , G r  a y

     G e o m e t  r   y

      S  E C O N

     D  E D I  T  I  O N

    This richly illustrated and clearly written undergraduate textbook captures

    the excitement and beauty of geometry. The approach is that of Klein

    in his Erlangen programme: a geometry is a space together with a set of

    transformations of the space. The authors explore various geometries:

    affine, projective, inversive, hyperbolic and elliptic. In each case they

    carefully explain the key results and discuss the relationships between the

    geometries.

      New features in this Second Edition include concise end-of-chapter

    summaries to aid student revision, a list of Further Reading and a list

    of Special Symbols. The authors have also revised many of the end-of-

    chapter exercises to make them more challenging and to include some

    interesting new results. Full solutions to the 200 problems are included

    in the text, while complete solutions to all of the end-of-chapter exercises

    are available in a new Instructors’ Manual , which can be downloaded from

    www.cambridge.org/9781107647831.

    Praise for the First Edition

    ‘To my mind, this is the best introductory book ever written on

    introductory university geometry… Not only are students introduced to a

    wide range of algebraic methods, but they will encounter a most pleasing

    combination of process and product.’

      P. N. RUANE , MAA Focus

    ‘… an excellent and precisely written textbook that should be studied in

    depth by all would-be mathematicians.’

      HANS SACHS,  American Mathematical Society

    ‘It conveys the beauty and excitement of the subject, avoiding the dryness

    of many geometry texts.’

      J . I. HALL , Michigan State University

    Geometry

    DAVI D A. BRA NNA

    MATTHEW F. ESPL

    JER EMY J. GRAY

    SECOND EDITI

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    2/602

    Geometry

    S E C O N D E D I T I O N

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    3/602

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    4/602

    GeometryS E C O N D E D I T I O N

    DAVI D A. BRAN NAN

    M AT T H E W F. E S P L E N

    J E R E M Y J . G R AY

    The Open University

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    5/602

    C A M B R I D G E U N I V E R S I T Y P R E S S

    Cambridge, New York, Melbourne, Madrid, Cape Town,Singapore, São Paulo, Delhi, Tokyo, Mexico City

    Cambridge University PressThe Edinburgh Building, Cambridge CB2 8RU, UK

    Published in the United States of America by Cambridge University Press, New York

    www.cambridge.orgInformation on this title: www.cambridge.org/9781107647831

    © The Open University 1999, 2012

    This publication is in copyright. Subject to statutory exceptionand to the provisions of relevant collective licensing agreements,no reproduction of any part may take place withoutthe written permission of Cambridge University Press.

    First published 1999Second edition 2012

    Printed in the United Kingdom at the University Press, Cambridge

     A catalogue record for this publication is available from the British Library

     Library of Congress Cataloguing in Publication data

    Brannan, D. A.Geometry / David A. Brannan, Matthew F. Esplen, Jeremy J. Gray. – 2nd ed.

    p. cm.ISBN 978-1-107-64783-1 (Paperback)

    1. Geometry. I. Esplen, Matthew F. II. Gray, Jeremy, 1947– III. Title.

    QA445.B688 2011516–dc23

    2011030683

    ISBN 978-1-107-64783-1 Paperback

    Additional resources for this publication at www.cambridge.org/9781107647831

    Cambridge University Press has no responsibility for the persistence oraccuracy of URLs for external or third-party internet websites referred toin this publication, and does not guarantee that any content on suchwebsites is, or will remain, accurate or appropriate.

    2.1

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    6/602

    In memory of Wilson Stothers

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    7/602

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    8/602

    Contents

    Preface page xi

    0 Introduction: Geometry and Geometries 1

    1 Conics 5

    1.1 Conic Sections and Conics 61.2 Properties of Conics 231.3 Recognizing Conics 361.4 Quadric Surfaces 421.5 Exercises 52

    Summary of Chapter 1 55

    2 Affine Geometry 61

    2.1 Geometry and Transformations 622.2 Affine Transformations and Parallel Projections 702.3 Properties of Affine Transformations 84

    2.4 Using the Fundamental Theorem of Affine Geometry 932.5 Affine Transformations and Conics 1082.6 Exercises 117

    Summary of Chapter 2 121

    3 Projective Geometry: Lines 127

    3.1 Perspective 1283.2 The Projective PlaneRP2 1373.3 Projective Transformations 1513.4 Using the Fundamental Theorem of Projective Geometry 1723.5 Cross-Ratio 1793.6 Exercises 192

    Summary of Chapter 3 195

    4 Projective Geometry: Conics 201

    4.1 Projective Conics 2024.2 Tangents 2164.3 Theorems 229

    vii

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    9/602

    viii   Contents

    4.4 Applying Linear Algebra to Projective Conics 2484.5 Duality and Projective Conics 2504.6 Exercises 252

    Summary of Chapter 4 255

    5 Inversive Geometry 261

    5.1 Inversion 2625.2 Extending the Plane 2765.3 Inversive Geometry 2955.4 Fundamental Theorem of Inversive Geometry 3105.5 Coaxal Families of Circles 3175.6 Exercises 331

    Summary of Chapter 5 335

    6 Hyperbolic Geometry: the Poincaŕe Model 343

    6.1 Hyperbolic Geometry: the Disc Model 3456.2 Hyperbolic Transformations 356

    6.3 Distance in Hyperbolic Geometry 3676.4 Geometrical Theorems 3836.5 Area 4016.6 Hyperbolic Geometry: the Half-Plane Model 4126.7 Exercises 413

    Summary of Chapter 6 417

    7 Elliptic Geometry: the Spherical Model 424

    7.1 Spherical Space 4257.2 Spherical Transformations 4297.3 Spherical Trigonometry 438

    7.4 Spherical Geometry and the Extended Complex Plane 4507.5 Planar Maps 4607.6 Exercises 464

    Summary of Chapter 7 465

    8 The Kleinian View of Geometry 470

    8.1 Affine Geometry 4708.2 Projective Reflections 4758.3 Hyperbolic Geometry and Projective Geometry 4778.4 Elliptic Geometry: the Spherical Model 4828.5 Euclidean Geometry 484

    Summary of Chapter 8 486

    Special Symbols   488

    Further Reading   490

     Appendix 1: A Primer of Group Theory   492

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    10/602

    Contents   ix

     Appendix 2: A Primer of Vectors and Vector Spaces   495

     Appendix 3: Solutions to the Problems   503Chapter 1 503Chapter 2 517Chapter 3 526

    Chapter 4 539Chapter 5 549Chapter 6 563Chapter 7 574

     Index    583

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    11/602

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    12/602

    Preface

    Geometry! For over two thousand years it was one of the criteria for recog-   Plato (c. 427–347 BC)was an Athenianphilosopher whoestablished a school of theoretical research (witha mathematical bias),

    legislation andgovernment.

    Archimedes (c. 287–212BC) was a Greekgeometer and physicistwho used many of thebasic limiting ideas of differential and integralcalculus.

    nition as an educated person to be acquainted with the subject of geometry.Euclidean geometry, of course.

    In the golden era of Greek civilization around 400 BC, geometry was studiedrigorously and put on a firm theoretical basis – for intellectual satisfaction, the

    intrinsic beauty of many geometrical results, and the utility of the subject.For example, it was written above the door of Plato’s Academy ‘Let no-oneignorant of Geometry enter here!’ Indeed, Archimedes is said to have used thereflection properties of a parabola to focus sunlight on the sails of the Romanfleet besieging Syracuse and set them on flame.

    For two millennia the children of those families sufficiently well-off to beeducated were compelled to have their minds trained in the noble art of rigor-ous mathematical thinking by the careful study of translations of the work of Euclid. This involved grasping the notions of axioms and postulates, the draw-   Euclid (c. 325–265 BC)

    was a mathematician inHellenistic Alexandriaduring the reign of Ptolemy I (323–283 BC),famous for his book The Elements.

    ing of suitable construction lines, and the careful deduction of the necessaryresults from the given facts and the Euclidean axioms – generally in two-

    dimensional or three-dimensional Euclidean space (which we shall denote byR2 and R3, respectively). Indeed, in the 1700s and 1800s popular publicationssuch as   The Lady’s and Gentleman’s Diary   published geometric problemsfor the consideration of gentlefolk at their leisure. And as late as the 1950s   We give a careful

    algebraic definition of R2

    and R3 in Appendix 2.translations of Euclid’s Elements were being used as standard school geometrytextbooks in many countries.

    Just as nowadays, not everyone enjoyed Mathematics! For instance, theGerman poet and philosopher Goethe wrote that ‘Mathematicians are like   Johann Wolfgang von

    Goethe (1749–1832) issaid to have studied allareas of science of his day

    except mathematics – forwhich he had no aptitude.

    Frenchmen: whatever you say to them, they translate into their own language,and forthwith it is something entirely different!’

    The Golden Era of geometry came to an end rather abruptly. When theUSSR launched the Sputnik satellite in 1957, the Western World suddenlydecided for political and military reasons to give increased priority to itsresearch and educational efforts in science and mathematics, and redevelopedthe curricula in these subjects. In order to make space for subjects newlydeveloped or perceived as more ‘relevant in the modern age’, the amount of geometry taught in schools and universities plummeted. Interest in geometrylanguished: it was thought ‘old-fashioned’ by the fashionable majority.

    xi

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    13/602

    xii   Preface

    Nowadays it is being realized that geometry is still a subject of abidingbeauty that provides tremendous intellectual satisfaction in return for effort putinto its study, and plays a key underlying role in the understanding, develop-ment and applications of many other branches of mathematics. More and moreuniversities are reintroducing courses in geometry, to give students a ‘feel’   Topics in computer

    graphics such as ‘hidden’

    surfaces and the shadingof curved surfaces involvemuch mathematics.

    for the reasons for studying various areas of mathematics (such as Topology),to service the needs of Computer Graphics courses, and so on. Geometry ishaving a revival!

    Since 1971, the Open University in the United Kingdom has taught math-ematics to students via specially written correspondence texts, and has tradi-tionally given geometry a central position in its courses. This book arises fromthose correspondence texts.

    We adopt the Klein approach to geometry. That is, we regard the variousgeometries as each consisting of an underlying set together with a group of transformations acting on that set. Those properties of the set that are notaltered by any of the transformations are called the properties of that geometry.

    Following a historical review of the development of the various geometries,   Chapter 0we look at conics (and at the related quadric surfaces) in Euclidean geometry.   Chapter 1Then we address a whole series of different geometries in turn. First, affine   Chapter 2geometry (that provides simple proofs of some results in Euclidean geome-try). Then projective geometry, which can be regarded as the most basic of    Chapters 3 and 4all geometries; we divide this material into a chapter on projective lines anda chapter on projective conics. We then return to study inversive geometry,   Chapter 5which provides beautiful proofs of many results involving lines and circles inEuclidean geometry. This leads naturally to the study of hyperbolic geometry   Chapter 6in the unit disc, in which there are two lines through any given point that areparallel to a given line. Via the link of stereographic projection, this leads on

    to spherical geometry: a natural enough concept for a human race that lives   Chapter 7on the surface of a sphere! Finally we tie things together, explaining how the   Chapter 8various geometries are inter-related.

    Study Guide 

    The book assumes a basic knowledge of Group Theory and of Linear Alge-bra, as these are used throughout. However, for completeness and students’convenience we give a very rapid review of both topics in the appendices.   Appendices 1 and 2.

    The book follows many of the standard teaching styles of The Open Uni-versity. Thus, most chapters are divided into five sections (each often furtherdivided into subsections); sections are numbered using two digits (such as‘Section 3.2’) and subsections using three digits (such as ‘Subsection 3.2.4’).Generally a section is considered to be about one evening’s hard work for anaverage student.

    We number in order the theorems, examples, problems and equations withineach section.

    We use wide pages with margins in which we place various historical notes,cross-references, teaching comments and diagrams; the cross-references need

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    14/602

    Preface   xiii

    not be consulted by students unless they wish to remind themselves of somepoint on that topic, but the other margin notes should be read carefully. We useboxes in the main text to highlight definitions, strategies, and the statements of theorems and other key results. The end of the proof of a theorem is indicatedby a solid symbol ‘’, and the end of the solution of a worked example by ahollow symbol ‘’. Occasionally the text includes a set of ‘Remarks’; these arecomments of the type that an instructor would give orally to a class, to clarifya definition, result, or whatever, and should be read carefully. There are manyworked examples within the text to explain the concepts being taught, and it isimportant that students read these carefully as they contain many key teachingpoints; in addition, there is a good stock of in-text problems to reinforce theteaching, and solutions to these are given in Appendix 3. At the end of eachchapter there are exercises covering the material of that chapter, some of whichare fairly straight-forward and some are more challenging; solutions are notgiven to the exercises.

    Our philosophy is to provide clear and complete explanations of all geomet-

    ric facts, and to teach these in such a way that students can understand themwithout much external help. As a result, students should be able to learn (and,we hope, to enjoy) the key concepts of the subject in an uncluttered way.

    Most students will have met many parts of Chapter 1 already, and so canproceed fairly quickly through it. Thereafter it is possible to tackle Chapters 2to 4 or Chapters 5 and 6, in either order. It is possible to omit Chapters 7 or 8,if the time in a course runs short.

    Notation for Functions as Mappings 

    Suppose that a function   f  maps some set   A into some set   B, and that it mapsa typical point x  of  A  onto some image point y of  B . Then we say that  A  is thedomain (or  domain of definition) of   f  ,   B  the codomain of   f  , and denote thefunction   f   as a  mapping (or map) as follows:   Note that we use two

    different arrows here, todistinguish between themapping of a set and themapping of an element.

     f   :   A →   B x  →  y

    We often denote   y  by the expression   f  ( x ) to indicate its dependence on   f and x .

    Acknowledgements 

    This material has been critically read by, or contributed to in some way,by many colleagues in The Open University and the BBC/OU TV Produc-tion Centre in Milton Keynes, including Andrew Adamyk, Alison Cadle,Anne-Marie Gallen, Ian Harrison, John Hodgson, Roy Knight, Alan Pears,Alan Slomson, Wilson Stothers and Robin Wilson. Its appearance in bookform owes a great deal to the work of Toni Cokayne, Pat Jeal and the OUMathematics and Computing Faculty’s Course Materials Production Unit.

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    15/602

    xiv   Preface

    Without the assistance and the forbearance of our families, the writing of the original OU course and its later rewriting in this form would have beenimpossible. It was Michael Brannan’s idea to produce it as a book.

    Changes in the Second Edition 

    In addition to correcting typos and errors, the authors have changed the term‘gradient’ to ‘slope’, and avoided the use of ‘reversed square brackets’ — sothat, for instance, the interval { x  : 0  <  x  ≤ 1} is now written as (0,1] ratherthan ]0,1]. Also, they have clarified the difference between a geometry and

    Solutions to the exercisesappear in an Instructors’Manual available from thepublisher.

    models of that geometry; in particular, the term ‘non-Euclidean’ geometry hasnow been largely replaced by ‘hyperbolic’ geometry, and the term ‘elliptic’geometry has been introduced where appropriate. The problems and exerciseshave been revised somewhat, and more exercises included. Each chapter nowincludes a summary of the material in that chapter, and before the appendicesthere are now lists of symbols and suggestions for further reading.

    The authors have taken the opportunity to add some new material to enrich

    the reader’s diet: a treatment of conics as envelopes of tangent families,barycentric coordinates, Poncelet’s Porism and Ptolemy’s Theorem, and planarmaps. Also, the treatment of a number of existing topics has been significantlychanged: the geometric interpretation of projective transformations, the anal-ysis of the formula for hyperbolic distance, and the treatment of asymptoticd -triangles.

    The authors appreciate the warm reception of the first edition, and have triedto take on board as many as possible of the helpful comments received. Specialthanks are due to John Snygg and Jonathan I. Hall for invaluable comments andadvice.

    Instructors’ Manual Complete solutions to all of the end-of-chapter exercises are available in anInstructors’ Manual, which can be downloaded from www.cambridge.org/ 9781107647831.

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    16/602

    0   Introduction: Geometryand Geometries

    Geometry is the study of shape. It takes its name from the Greek belief that   The word comes from theGreek words geo (Earth)and metria (measuring).

    geometry began with Egyptian surveyors of two or three millennia ago mea-suring the Earth, or at least the fertile expanse of it that was annually floodedby the Nile.

    It rapidly became more ambitious. Classical Greek geometry, called Euclidean geometry after Euclid, who organized an extensive collection of theorems into his definitive text The Elements, was regarded by all in the early   Isaac Newton

    (1643–1727) was anEnglish astronomer,physicist andmathematician. He wasProfessor of Mathematicsat Cambridge, Master of the Royal Mint, andsuccessor of SamuelPepys as President of theRoyal Society.

    modern world as the true geometry of space. Isaac Newton used it to formu-late his  Principia Mathematica  (1687), the book that first set out the theoryof gravity. Until the mid-19th Century, Euclidean geometry was regarded asone of the highest points of rational thought, as a foundation for practicalmathematics as well as advanced science, and as a logical system splendidlyadapted for the training of the mind. We shall see in this book that by the 1850sgeometry had evolved considerably – indeed, whole new geometries had beendiscovered.

    The idea of using coordinates in geometry can be traced back to Apollo-

    nius’s treatment of conic sections, written a generation after Euclid. But their   Apollonius of Perga(c. 255–170 BC) was aGreek geometer, whoseonly surviving work is atext on conics.

    use in a systematic way with a view to simplifying the treatment of geome-try is really due to Fermat and Descartes. Fermat showed how to obtain anequation in two variables to describe a conic or a straight line in 1636, but hiswork was only published posthumously in 1679. Meanwhile in 1637 Descartespublished his book  Discourse on Method , with an extensive appendix enti-   Pierre de Fermat

    (1601–1665) was a Frenchlawyer and amateurmathematician, whoclaimed to have a proof of the recently provedFermat’s Last Theorem in

    Number Theory.

    tled La Géometrie, in which he showed how to introduce coordinates to solvea wide variety of geometrical problems; this idea has become so central apart of mathematics that whole sections of  La Géometrie read like a moderntextbook.

    A contemporary of Descartes, Girard Desargues, was interested in the ideas

    René Descartes(1596–1650) was a Frenchscientist, philosopher andmathematician. He is alsoknown for the phrase‘Cogito, ergo sum’ (Ithink, therefore I am).

    of perspective that had been developed over many centuries by artists (anx-ious to portray three-dimensional scenes in a realistic way on two-dimensionalwalls or canvases). For instance, how do you draw a picture of a building,or a staircase, which your client can understand and commission, and fromwhich artisans can deduce the correct dimensions of each stone? Desarguesalso realized that since any two conics can always be obtained as sections of the same cone in R3, it is possible to present the theory of conics in a unified

    1

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    17/602

    2   0: Introduction: Geometry and Geometries

    way, using concepts which later mathematicians distilled into the notion of    Girard Desargues(1591–1661) was a Frencharchitect.

    the cross-ratio of four points. Desargues’ discoveries came to be known as projective geometry.

    We deal with these ideasin Chapters 4 and 5.Blaise Pascal was the son of a mathematician, Étienne, who attended a group

    of scholars frequented by Desargues. He heard of Desargues’s work from hisfather, and quickly came up with one of the most famous results in the geom-etry of conics, Pascal’s Theorem, which we discuss in Chapter 4. By the late   Blaise Pascal

    (1623–1662) was a Frenchgeometer, probabilist,physicist and philosopher.

    19th century projective geometry came to be seen as the most basic geometry,with Euclidean geometry as a significant but special case.

    At the start of the 19th century the world of mathematics began to change.The French Revolution saw the creation of the École Polytechnique in Parisin 1794, an entirely new kind of institution for the training of military engi-neers. It was staffed by mathematicians of the highest calibre, and run for   Gaspard Monge

    (1746–1818) was a Frenchanalyst and geometer. Astrong republican andsupporter of the

    Revolution, he was FrenchMinister of the Navy in1792–93, but deprived of all his honours on therestoration of the Frenchmonarchy.

    many years by Gaspard Monge, an enthusiastic geometer who had invented asimple system of  descriptive geometry for the design of forts and other mili-tary sites. Monge was one of those rare teachers who get students to see what is

    going on, and he inspired a generation of French geometers. The École Poly-technique, moreover, was the sole entry-point for any one seeking a careerin engineering in France, and the stranglehold of the mathematicians ensuredthat all students received a good, rigorous education in mathematics beforeentering the specialist engineering schools. Thus prepared they then assistedNapoleon’s armies everywhere across Europe and into Egypt.

    One of the École’s former students, Jean Victor Poncelet, was taken prisoner   Jean Victor Poncelet(1788–1867) followed acareer as a militaryengineer by becomingProfessor of Mechanics atMetz, where he worked on

    the efficiency of turbines.

    in 1812 in Napoleon’s retreat from Moscow. He kept his spirits up during aterrible winter by reviewing what his old teacher, Monge, had taught him aboutdescriptive geometry. This is a system of projections of a solid onto a plane –or rather two projections, one vertically and one horizontally (giving what are

    called to this day the  plan and  elevation of the solid). Poncelet realized thatinstead of projecting ‘from infinity’ so to speak, one could adapt Monge’s ideas

    vertical

    projection

    horizontal

    projection

    elevation

    plan

    21

    to the study of projection from a point. In this way he re-discovered Desargues’ideas of projective geometry. During his imprisonment he wrote his famousbook Traité des propriétés projectives des figures outlining the foundations of projective geometry, which he extensively rewrote after his release in 1814 andpublished in 1822.

    Around the same time that projective geometry was emerging, mathemati-cians began to realize that there was more to be said about circles than theyhad previously thought. For instance, in the study of electrostatics let  1 and2 be two infinitely long parallel cylinders of opposite charge. Then the inter-section of the surfaces of equipotential with a vertical plane is two families of circles (and a single line), and a point charge placed in the electrostatic fieldmoves along a circular path through a specific point inside each cylinder, atright angles to circles in the families. The study of properties of such fami-lies of circles gave rise to a new geometry, called  inversive geometry, whichwas able to provide particularly striking proofs of previously known results inEuclidean geometry as well as new results.

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    18/602

    0: Introduction: Geometry and Geometries   3

    In inversive geometry mathematicians had to add a ‘point at infinity’ to theplane, and had to regard circles and straight lines as equivalent figures underthe natural mappings, inversions, as these can turn circles into lines, and vice-versa. Analogously, in projective geometry mathematicians had to add a whole‘line at infinity’ in order to simplify the geometry, and found that there wereprojective transformations that turned hyperbolas into ellipses, and so on. Somathematicians began to move towards thinking of geometry as the study of shapes and the transformations that preserve (at least specified properties of)those shapes.

    For example, there are very few theorems in Euclidean geometry that dependon the size of the figure. The ability to make scale copies without altering‘anything important’ is basic to mathematical modelling and a familiar factof everyday life. If we wish to restrict our attention to the transformationsthat preserve length, we deal with Euclidean geometry, whereas if we allowarbitrary changes of scale we deal with  similarity geometry.

    Another interesting geometry was discovered by Möbius in the 1820s, in   August Ferdinand Möbius(1790–1868) was aGerman geometer,topologist, numbertheorist and astronomer;he discovered the famousMöbius Strip (or Band).

    which transformations of the plane map lines to lines, parallel lines to parallellines, and preserve ratios of lengths along lines. He called this geometry  affinegeometry because any two figures related by such a transformation have a like-ness or affinity to one another. This is the geometry appropriate, in a sense, toMonge’s descriptive geometry, and the geometry that describes the shadows of figures in sunlight.

    Since the days of Greek mathematics, with a stimulus provided by the needs   For the surface of theEarth is very nearlyspherical.

    of commercial navigation, mathematicians had studied spherical geometry too;that is, the geometry of figures on the surface of a sphere. Here geometry

    a

    b

    cis rather different from plane Euclidean geometry; for instance the area of a triangle is proportional to the amount by which its angle sum exceeds   π ,

    and there is a nice generalization of Pythagoras’ Theorem, which says thatin a right-angled triangle with sides  a , b and the hypotenuse  c, then cos c =cos a · cos b. It turns out that there is a close connection between sphericalgeometry and inversive geometry.

    For nearly two millennia mathematicians had accepted as obvious the   pm   lParallel Postulate of Euclid: namely, that given any line    and any point   P

    not on  , there is a unique line  m  in the same plane as   P  and   which passesthrough   P  and does not meet  . Indeed much effort had been put into deter-mining whether this Postulate could be deduced from the other assumptions   Janos Bolyai (1802–1860)

    was an officer in theHungarian Army.

    of Euclidean geometry. In the 1820s two young and little-known mathemati-cians, Bolyai in Hungary and Lobachevskii in Russia, showed that therewere perfectly good so-called ‘non-Euclidean geometries’, namely   hyper-bolic geometry and  elliptic geometry, that share all the initial assumptions of    Nicolai Ivanovich

    Lobachevskii(1792–1856) was aRussian geometer whobecame Rector of theUniversity of Kazan.

    Euclidean geometry except the parallel postulate.In hyperbolic geometry given any line  and any point  P not on , there are

    infinitely many lines in the same plane as  P  and which pass through  P  and donot meet ; in elliptic geometry all lines intersect each other. However, it stillmakes sense in both hyperbolic and elliptic geometries to talk about the length

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    19/602

    4   0: Introduction: Geometry and Geometries

    of line segments, the distance between points, the angles between lines, andso forth. Around 1900 Poincaré did a great deal to popularise these geometries   Jules Henri Poincaré

    (1854–1912) was aprolific Frenchmathematician, physicist,astronomer and

    philosopher at theUniversity of Paris.

    by demonstrating their applications in many surprising areas of mathematics,such as Analysis.

    By 1870, the situation was that there were many geometries: Euclidean,affine, projective, inversive, hyperbolic and elliptic geometries. One way math-ematicians have of coping with the growth of their subject is to re-define it sothat different branches of it become branches of the same subject. This wasdone for geometry by Klein, who developed a programme (the  Erlangen Pro-   Christian Felix Klein

    (1849–1925) was aGerman algebraist,geometer, topologist andphysicist; he became aprofessor at the Universityof Erlangen at theremarkable age of 22.

    gramme) for classifying geometries. His elegant idea was to regard a  geometryas a space together with a group of transformations of that space; the proper-ties of figures that are not altered by any transformation in the group are theirgeometrical properties.

    For example, in two-dimensional Euclidean geometry the space is the planeand the group is the group of all length-preserving transformations of the plane(or isometries). In projective geometry the space is the plane enlarged (in a way

    we make precise in Chapter 6) by a line of extra points, and the group is thegroup of all continuous transformations of the space that preserve  cross-ratio.Klein’s approach to a geometry involves three components: a set of points

    (the space), a set of transformations (that specify the invariant properties – forexample, congruence in Euclidean geometry), and a group (that specifies howthe transformations may be composed). The transformations and their groupare the fundamental components of the geometry that may be applied to differ-ent spaces. A model of a geometry is a set which possesses all the properties of the geometry; two different models of any geometry will be isomorphic. Theremay be several different models of a given geometry, which have different   For example, you will

    meet two models of 

    hyperbolic geometry.

    advantages and disadvantages. Therefore, we shall use the terms ‘geometry’

    and ‘model (of a geometry)’ interchangeably whenever we think that there isno risk of confusion.

    In fact as Klein was keen to stress, most geometries are examples of pro- jective geometry with some extra conditions. For example, affine geometryemerges as the geometry obtained from projective geometry by selecting a lineand considering only those transformations that map that line to itself; the linecan then be thought of as lying ‘at infinity’ and safely ignored. The result wasthat Klein not only had a real insight into the nature of geometry, he could evenshow that projective geometry was almost the most basic geometry.

    This philosophy of geometry, called the  Kleinian view of geometry, is theone we have adopted in this book. We hope that you will enjoy this introductionto the various geometries that it contains, and go on to further study of one of the oldest, and yet most fertile, branches of mathematics.

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    20/602

    1   Conics

    The study of conics is well over 2000 years old, and has given rise to some of the most beautiful and striking results in the whole of geometry.

    In Section 1.1 we outline the Greek idea of a  conic section – that is, a conicas defined by the curve in which a double cone is intersected by a plane. Wethen look at some properties of circles, the simplest of the non-degenerateconics, such as the condition for two circles to be  orthogonal and the equations   That is, they intersect at

    right angles.of the family of all circles through two given points.We explain the focus–directrix definition of the parabola, ellipse and hyper-

    bola, and study the focal-distance properties of the ellipse and hyperbola.Finally, we use the so-called  Dandelin spheres to show that the Greek conicsections are just the same as the conics defined in terms of a focus and adirectrix.

    In Section 1.2 we look at tangents to conics, and the reflection properties of the parabola, ellipse and hyperbola. It turns out that these are useful in prac-tical situations as diverse as anti-aircraft searchlights and astronomical opticaltelescopes! We also see how we can construct each non-degenerate conic as

    the ‘envelope’ of lines in a suitably-chosen family of lines.The equations of conics are all second degree equations in   x   and   y. In

    Section 1.3 we show that the converse result holds – that is, that every sec-ond degree equation in  x  and   y represents a conic. We also find an algorithmfor determining from its equation in   x   and   y  which type of non-degenerateconic a given second degree equation represents, and for finding its principalfeatures.

    The analogue in  R3 of a plane conic in  R2 is a  quadric surface, specified   We use the notation R2

    and R3 to denote2-dimensional and3-dimensional Euclidean

    space, respectively.

    by a suitable second degree equation in   x , y  and   z. A well-known exampleof a quadric surface is the cooling tower of an electricity generating station.In Section 1.4 we find an algorithm for identifying from its equation which

    type of non-degenerate quadric a given second degree equation in  x, y and  zrepresents. We also discover that two of the non-degenerate quadric surfacescan be generated by two different families of straight lines, and that this featureis of practical importance.

    5

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    21/602

    6   1: Conics

    1.1 Conic Sections and Conics

    1.1.1 Conic Sections

    Conic Section is the name given to the shapes that we obtain by taking different

    It is thought that theGreek mathematicianMenaechmus discoveredthe conic sections around

    350 BC.

    plane slices through a double cone. The shapes that we obtain from these cross-

    sections are as drawn below.

    6

    5

    1. single point 2. single line 3. pair of lines

    O

    O O O

    O O x 

     x x x 

     x x 

     y

     y y y

     y y

    1

    3

    7

    6 4 2

    4. parabola 5. ellipse 6. hyperbola

    Notice that the circle shown in slice 7 can be regarded as a special case of 

     y

    O

    7. Circle

     x an ellipse.

    Notice, also, that the ellipse and the hyperbola both have a   centre; that is,there is a point C  such that rotation about C  through an angle π  is a symmetryof the conic. For example, for the ellipse and hyperbola illustrated above, thecentre is in fact just the origin. On the other hand, the parabola does not have

    a centre.In Subsection 1.1.5 we shall verify that the curves, the ‘conic sections’,

    obtained by slicing through a double cone are exactly the same curves, the‘conics’, obtained as the locus of points in the plane whose distance froma fixed point is a constant multiple of its distance from a fixed line. As aresult, we often choose not to distinguish between the terms ‘conic section’and ‘conic’!

    We use the term   non-degenerate conics  to describe those   conics  that areparabolas, ellipses or hyperbolas; and the term  degenerate conics to describethe single point, single line and pair of lines.

    In this chapter we study conics for their own interest, and we will meet themfrequently throughout our study of geometry as the book progresses.

    1.1.2 Circles

    The first conic that we investigate is the circle. Recall that a  circle in R2 is theC (a, b)

     y

    P( x , y)

     x 

    set of points ( x , y) that lie at a fixed distance, called the  radius, from a fixedpoint, called the  centre of the circle. We can use the techniques of coordinategeometry to find the equation of a circle with given centre and radius.

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    22/602

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    23/602

    8   1: Conics

    For example, consider the set of points  ( x , y) in the plane that satisfy the

    Note that in equation (3)the coefficients of  x 2 and y2 are both 1.

    equation:

     x 2 +  y 2 − 4 x  + 6 y + 9 = 0. (3)In order to transform equation (3) into an equation of the form (1), we use the

    technique called ‘completing the square’

     − we rewrite the terms that involve

    only x s and the terms that involve only  y s as follows:

    Note that −2 is half thecoefficient of  x , and +3 ishalf the coefficient of  y , inequation (3).

     x 2 − 4 x  =  ( x  − 2)2 − 4, y2 + 6 y = ( y + 3)2 − 9.

    Substituting these expressions into equation (3), we obtain

    We can ‘read off’ thecentre and radius of thecircle from this equation.

    ( x  − 2)2 + ( y + 3)2 = 4.

    It follows that the equation represents a circle whose centre is (2, −3) andwhose radius is 2.

    In general, we can use the same method of ‘completing the square’ to rewrite   Here we start with thecoefficients of  x 2 and  y 2both equal (to 1).Otherwise the equationcannot be reformulated inthe form (1).

    the equation

     x 2 +  y 2 + fx + gy + h = 0in the form

     x  +   12   f 2

    +

     y +  12 g2

    =   14   f  2 +   14 g2 − h, (4)from which we can ‘read off’ the centre and radius.

    Theorem 2   An equation of the form

     x 2

    + y 2

    + fx 

    +gy

    +h

     = 0

    represents a circle with

    centre− 12   f  , − 12 g

     and radius

     14   f 

     2 +   14 g2 − h,

    provided that   14   f  2 +  14 g2 − h  > 0.

    Remark 

    It follows from equation (4) above that if   14   f  2 +   14 g2 − h  

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    24/602

    Conic Sections and Conics   9

    Problem 4   Determine the set of points ( x , y) in R2 that satisfies eachof the following equations:

    (a)   x 2 +  y 2 +  x  + y + 1 = 0;(b)   x 2 +  y 2 − 2 x  + 4 y + 5 = 0;(c) 2 x 2

    +2 y2

    + x 

     −3 y

     −5

     = 0.

    Orthogonal Circles 

    We shall sometimes be interested in whether two intersecting circles are   For example, inChapters 5 and 6.orthogonal: that is, whether they meet at right angles. The following result

    answers this question if we know the equations of the two circles.

    Theorem 3 Orthogonality Test

    Two intersecting circles C 1 and  C 2 with equations

     x 2

    + y 2

    +  f 1 x 

     +g1 y

     +h1

     = 0 and

     x 2 + y 2 +   f 2 x  + g2 y + h2 = 0,respectively, are orthogonal if and only if 

     f 1 f 2 + g1g2 = 2(h1 + h2).

    Proof    The circle   C 1   has centre   A   =− 12   f 1, − 12 g1

      and radius   r 1   =   You met these formulas in

    Theorem 2. 14   f 

     21 + 14 g21 − h1; the circle  C 2  has centre   B =

    − 12   f 2, − 12 g2

     and radius

    r 2 =  14   f 

     22 + 12 g22 − h2.

    Let  P be one of their points of intersection, and look at the triangle   ABP.   We use the symbol   toindicate a triangle.If the circles meet at right angles, then the line  AP is tangential to the circleC 2, and is therefore at right angles to the line   BP. So the triangle   ABP is

     A

    C 1   C 2

    P

     B

    right-angled, and we may apply Pythagoras’ Theorem to it to obtain

     AP2 + BP2 =  AB2. (5)

    Conversely, if equation (5) holds, then  ABP must be a right-angled triangleand the circles must meet at right angles.

    Now

     AP2

    = r 21

     =  14   f 

     21

     +  14 g

    21

     −h1   and

     BP2 = r 22 =   14   f  22 +  14 g22 − h2.

    Also

     AB2 =

    12   f 1 −  12   f 2

    2+

    12 g1 −  12 g2

    2=

    14   f 

     21 −  12   f 1 f 2 +  14   f  22

    +

    14 g

    21 −  12 g1g2 + 14 g22

    .

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    25/602

    10   1: Conics

    Substituting for   AP2,   BP2 and   AB2 into equation (5), and cancellingcommon terms, we deduce that equation (5) is equivalent to

    −h1 − h2 = −12   f 1 f 2 −  12 g1g2,that is,

     f 1 f 2

     +g1g2

     = 2(h1

     +h2).

    This is the required result.  

    Problem 5   Determine which, if any, of the following pairs of inter-secting circles are mutually orthogonal.

    (a)   C 1 =

    ( x , y) :  x 2 +  y 2 − 4 x  − 4 y + 7 = 0   andC 2 =

    ( x , y) :  x 2 +  y 2 + 2 x  − 8 y + 5 = 0

    (b)   C 1 =

    ( x , y) :  x 2 +  y 2 + 3 x  − 6 y + 5 = 0   andC 2 =

    ( x , y) : 3 x 2 + 3 y2 + 4 x  + y − 15 = 0.

    Circles through Two Points 

    We shall also be interested later in the family of circles through two given   Section 5.5points. So, let two circles C 1 and  C 2 with equations

    P

     R

    C 1

    C 2

    Q

     x 2 +  y 2 +   f 1 x  + g1 y + h1 = 0 and x 2 +  y 2 +   f 2 x  + g2 y + h2 = 0 (6)

    intersect at the distinct points  P and  Q, say. Then, if  k  = −1, the equation x 2 + y 2 +   f 1 x  + g1 y + h1 + k ( x 2 + y 2 +   f 2 x  + g2 y + h2) = 0 (7)

    represents a circle since it is a second degree equation in  x  and   y with equal(non-zero) coefficients of   x 2 and   y2 and with no terms in   xy. This circle

    passes through both   P   and   Q; for the coordinates of   P   and   Q  both satisfythe equations in (6) and so must satisfy equation (7).

    If  k  = −1, equation (7) is linear in  x  and  y, and so represents a line; sinceP and  Q both lie on it, it must be the line through  P and  Q .

    Conversely, given any point   R  in the plane that does not lie on the circleC 2 we can substitute the coordinates of   R into equation (7) to find the unique   This is possible because,

    since R  does not lie onC 2, the term in the bracketin (7) does not vanishat  R.

    value of  k  such that the circle with equation (7) passes through  R . We can thinkof the circle C 2 as corresponding to the case ‘k  = ∞’ of equation (7). For, if we rewrite equation (7) in the form

    1

    ( x 2

    + y 2

    +  f 1 x 

     +g1 y

     +h1)

    + x 2

    + y 2

    +  f 2 x 

     +g2 y

     +h2

     = 0 (8)

    and let k  → ∞, then 1/k  → 0 and equation (8) becomes the equation of  C 2.

    Theorem 4   Let C 1 and  C 2 be circles with equations

     x 2 +  y 2 +   f 1 x  + g1 y + h1 = 0 and x 2 +  y 2 +   f 2 x  + g2 y + h2 = 0

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    26/602

    Conic Sections and Conics   11

    that intersect at distinct points  P and  Q . Then the line and all circles (otherthan C 2) through  P and  Q  have an equation of the form

     x 2 +  y 2 +   f 1 x  + g1 y + h1 + k ( x 2 +  y 2 +   f 2 x  + g2 y + h2) = 0for some number k .

    If  k  = −1, this equation is one of the circles; if   k  = −1, this is theequation of the line.

    Example 1   Find the equation of the circle that passes through (1, 2) and thepoints of intersection of the circles

     x 2 +  y 2 − 3 x  + 4 y − 1 = 0 and   x 2 + y 2 +   52 x  − 3 y +  32 = 0.

    Solution    By Theorem 4, the required equation is of the form

     x 2 +  y 2 − 3 x  + 4 y − 1 + k 

     x 2 +  y 2 +  52 x  − 3 y +  32 = 0 (9)

    for some number k . Since (1, 2) must satisfy this equation, it follows that

    1 + 4 − 3 + 8 − 1 + k 

    1 + 4 +   52 − 6 +   32

    = 0,so that k  = −3. Substituting k  = −3 back into equation (9), we deduce thatthe equation of the required circle is

     x 2 +  y 2 − 3 x  + 4 y − 1 − 3

     x 2 +  y 2 +   52 x  − 3 y +  32

    = 0,which we can simplify to the form

    4 x 2 + 4 y2 + 21 x  − 26 y + 11 = 0.

    Problem 6   Find the equation of the line through the points of inter-section of the circles

     x 2 +  y 2 − 3 x  + 4 y − 1 = 0 and2 x 2 + 2 y2 + 5 x  − 6 y + 3 = 0.

    1.1.3 Focus-Directrix Definition of the Non-Degenerate ConicsEarlier we defined the conic sections as the curves of intersection of a double   Subsection 1.1.1

    cone with a plane. We have seen that the circle can be defined in a different   Subsection 1.1.2way: as the set of points at a fixed distance from a fixed point.

    Here we give a method for constructing the other non-degenerate conics,the parabola, ellipse and hyperbola, as sets of points that satisfy a   some-what  similar condition involving distances. Later we shall give a careful proof    Subsection 1.1.5

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    27/602

    12   1: Conics

    that each non-degenerate conic section is a non-degenerate (plane) conic,and vice-versa.   Theorem 4 of 

    Subsection 4.1.4The three  non-degenerate conics (the parabola, ellipse and hyperbola) canbe defined as the set of points   P  in the plane that satisfy the following condi-tion: The distance of   P  from a fixed point (called the  focus of the conic) is aconstant multiple (called its  eccentricity, e) of the distance of   P  from a fixedline (called its directrix ).

    The different conics arise according to the value of the eccentricity:

    Eccentricity   A non-degenerate conic is an ellipse if 0 ≤ e  <  1, a parabola   When e = 0, the ellipse isactually a circle; the focusis the centre of the circle,and the directrix is ‘atinfinity’.

    if  e = 1, or a hyperbola if  e  >  1.

    Parabola ( e = 1) A   parabola is defined to be the set of points   P  in the plane whose distance

     M 

    O

     x  = –a

    directrix d 

     x 

    focus

    F  (a, 0)

     y   FP = PM 

    P ( x , y)from a fixed point  F  is equal to their distance from a fixed line  d . We obtain a

    parabola in standard form if we choose

    1. the focus  F  to lie on the  x -axis, and to have coordinates (a, 0), a  > 0;2. the directrix d  to be the line with equation  x  = −a.

    Notice in particular that the origin   O(0, 0)   lies on the parabola since it isequidistant from  F  and  d .

    Let   P( x , y) be an arbitrary point on the parabola, and let   M  be the foot of the perpendicular from   P  to the directrix. Since  FP =   PM , by the definitionof the parabola, it follows that  FP2 =   PM 2; we may rewrite this equation interms of coordinates as

    ( x  − a)2 +  y 2 =  ( x  + a)2.

    Multiplying out the brackets we get

     x 2 − 2ax + a2 +  y 2 =  x 2 + 2ax + a2,

    which simplifies to the equation  y 2 = 4ax .Notice that each point with coordinates

    at 2, 2at 

    , where t  ∈  R, lies on the   We use the notation R to

    denote the ‘real numbers’or the ‘real line’.

    parabola, since (2at )2 = 4a · at 2. Conversely, we can write the coordinates of each point on the parabola in the form

    at 2, 2at 

    . For if we choose t  =  y/(2a),

    then  y

     = 2at  and

     x  =   y2

    4a(from the equation  y 2 = 4ax )

    =   (2at )2

    4a=  at 2,

    as required. It follows that there is a one–one correspondence between the realnumbers t  and the points of the parabola.

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    28/602

    Conic Sections and Conics   13

    We summarize the above facts as follows.

    Parabola in Standard Form   A parabola in standard form has equation

     y2 = 4ax , where a  >  0.It has focus  (a, 0) and directrix   x 

     = −a; and it can be described by the

    parametric equations

     x  =  at 2,   y = 2at    (t  ∈ R).

    We call the   x -axis the   axis  of the parabola in standard form, since theparabola is symmetric with respect to this line, and we call the origin the vertex of a parabola in standard form, since it is the point of intersection of the axis   ‘Centre’ was defined in

    Subsection 1.1.1.of the parabola with the parabola. A parabola has no centre.

    Example 2   This question concerns the parabola E  with equation y 2 = 2 x  and   We generally use the letter E  to denote a conic.parametric equations x  =   12 t 2,  y = t  (t  ∈ R).

    (a) Write down the focus, vertex, axis and directrix of  E .(b) Determine the equation of the chord that joins distinct points  P and  Q  on

     E  with parameters  t 1 and  t 2, respectively. Determine the condition on   t 1and t 2 such that the chord  PQ passes through the focus of  E .   Such a chord is called

     focal chord.

    Solution 

    (a) The parabola  E  is the parabola in standard form where 4a = 2, or a =   12 . y

     x F  (1–2, 0)

    Q (1–2 t 22, t 2)

    P (1–2 t 21, t 1)

    It follows that the focus of   E   is

    12 , 0

    , its vertex is (0, 0), its axis is the x -axis, and the equation of its directrix is x  = −12 .

    (b) The coordinates of  P and  Q are

    12 t 

    21, t 1

    and

    12 t 

    22, t 2

    , respectively. So,

    if   t 21  =   t 22, the   slope   (or   gradient , as it is sometimes called) of   PQ   isgiven by

    m =   t 1 − t 212 t 

    21 − 12 t 22

    =   t 1 − t 212

    t 21 − t 22

     =   2t 1 + t 2

    .

    Since

    12 t 

    21, t 1

    lies on the line PQ, it follows that the equation of  PQ is

     y − t 1 =  2

    t 1 + t 2

     x  − 1

    2 t 21

    .

    Multiplying both sides by  t 1 + t 2, we get(t 1 + t 2)( y − t 1) = 2 x  − t 21,

    so that

    (t 1 + t 2) y − t 21 − t 1t 2 = 2 x  − t 21,

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    29/602

    14   1: Conics

    or y

     E 

    P (t 21,t 1)

    Q (t 22,t 2)

     x 

     y2 = x (t 1 + t 2) y = 2 x  + t 1t 2. (10)If, however,   t 21 =   t 22, then since  t 1 =   t 2  we have   t 1 = −t 2. Thus  PQ isparallel to the  y -axis, and so has equation  x  =   12 t 21; so in this case too,  PQhas equation given by (10).

    The chord   PQ  with equation (10) passes through the focus 1

    2 , 0

      if (t 1 + t 2)0 = 1 + t 1t 2; in other words, if  t 1t 2 = −1.

    Problem 7   This question concerns the parabola   E   with equation y2 =  x  and parametric equations x  =  t 2,  y = t  (t  ∈ R).(a) Write down the focus, vertex, axis and directrix of  E .(b) Determine the equation of the chord that joins distinct points   P and

    Q on  E  with parameters t 1 and  t 2, respectively.(c) Determine the condition on  t 1 and t 2  (and so on   P  and   Q) that the

    focus of  E  is the midpoint of the chord  PQ.

    Ellipse ( 0 ≤ e  <  1) We define an ellipse   with eccentricity zero   to be a circle. We have already   Subsection 1.1.2discussed circles.

    We define an  ellipse with eccentricity  e (where 0   <   e   <  1) to be the set y

     M P ( x , y)

    F  (ae, 0)focus

     x  = a / e

    directrix

     x O

    of points   P  in the plane whose distance from a fixed point   F   is e  times theirdistance from a fixed line  d . We obtain such an ellipse  in standard form if wechoose

    1. the focus  F  to lie on the  x -axis, and to have coordinates (ae, 0), a  >  0;2. the directrix d  to be the line with equation  x  =  a/e.

    Let  P( x , y) be an arbitrary point on the ellipse, and let  M  be the foot of theperpendicular from  P to the directrix. Since  FP = e · PM , by the definition of the ellipse, it follows that  FP2 =   e2 · PM 2; we may rewrite this equation interms of coordinates as

    ( x  − ae)2 +  y 2 = e2

     x  −  ae

    2= (ex − a)2.

    Multiplying out the brackets we get

     x 2

    −2aex 

    +a2e2

    + y 2

    = e2 x 2

    −2aex 

    +a2,

    which simplifies to the equation

     x 2

    1 − e2

    +  y 2 =  a2

    1 − e2

    or x 2

    a2 +   y

    2

    a21 − e2 = 1.

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    30/602

    Conic Sections and Conics   15

    Substituting   b   for   a√ 

    1 − e2, so that   b2 =   a2 1 − e2, we obtain the   Since 0  <  e  <  1, we havethat 0 <  b  <  a .standard form of the equation of the ellipse

     y

    b

    focus

    directrix

     x  = –a / e

    directrix

     x  = a / e

    focus

    minor axis

    major axis centre

    F   a   x F ′ –a

     –b

     x 2

    a2 +   y

    2

    b2 = 1.

    Notice that this equation is symmetrical in x 

     and symmetrical in  y

    , so thatthe ellipse also has a second focus   F (−ae, 0) and a second directrix  d  withequation x  = −a/e.

    The ellipse intersects the axes at the points   (±a, 0) and   (0, ±b). We callthe segment joining the points  (±a, 0) the  major axis of the ellipse, and thesegment joining the points  (0, ±b) the  minor axis of the ellipse. Since  b   <  a ,the minor axis is shorter than the major axis. The origin is the centre of this   Sometimes it is

    convenient to assume thatt  ∈ [0, 2π ), for instance,instead of  (−π , π ]. Noticeour notation for intervals:

    ( p, q), [ p, q], [ p, q), ( p, q]denote those real numbers x  for which   p  <  x    0, it has foci  (±ae, 0) and directrices x  = ±a/e.

    Example 3   Let PQ be an arbitrary chord of the ellipse with equation y P

    O

     x 2 / a

    2 + y2 / b

    2 = 1

     M 

    Q x 

     x 2

    a2 +   y

    2

    b2 = 1.

    Let   M   be the midpoint of   PQ. Prove that the following expression isindependent of the choice of  P and  Q :

    slope of  OM × slope of  PQ.

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    31/602

    16   1: Conics

    Solution    Let   P   and   Q   have the parametric coordinates   (a cos t 1, b sin t 1)and   (a cos t 2, b sin t 2), respectively. It follows that   M    has coordinates( a2 (cos t 1 + cos t 2),   b2 (sin t 1 + sin t 2)).

    Now,

    the slope of  OM  =   b(sin t 1 + sin t 2)a(cos t 1

     +cos t 2)

    andthe slope of  PQ =   b(sin t 1 − sin t 2)

    a(cos t 1 − cos t 2),

    so

    In general,

    cos2 θ  = 1 − sin2 θ .

    slope of  OM × slope of  PQ

    =   b(sin t 1 + sin t 2)a(cos t 1 + cos t 2)

    ·   b(sin t 1 − sin t 2)a(cos t 1 − cos t 2)

    =   b2

    a2 ·   sin

    2 t 1 − sin2 t 2cos2 t 1 − cos2 t 2

    =  b2

    a2 ·  sin2 t 1

     −sin2 t 2

    1 − sin2 t 1− 1 − sin2 t 2

    = −b2

    a2,

    which is independent of the values of  t 1 and  t 2.

    Problem 8   Let   P  be an arbitrary point on the ellipse with equation   y

     x 

     x 2 / a

    2 + y2 / b

    2 = 1

     M 

    P (a cos t, b sin t )

    F  (ae, 0)

     x 2

    a2 +  y2

    b2 = 1 and focus  F (ae, 0). Let M  be the midpoint of  FP. Prove that

     M  lies on an ellipse whose centre is midway between the origin and  F .

    Hyperbola ( e  >  1) 

    A  hyperbola is the set of points   P  in the plane whose distance from a fixed

     y

     x 

     M P( x , y)

    F (ae, 0)

    focus

     x  = a/e

    directrix

    O

    point F  is  e times their distance from a fixed line d , where e  >  1. We obtain ahyperbola in standard form if we choose

    1. the focus  F  to lie on the  x -axis, and to have coordinates (ae, 0), a  >  0;2. the directrix d  to be the line with equation  x  =  a/e.

    Let  P( x , y) be an arbitrary point on the hyperbola, and let  M  be the foot of the perpendicular from  P  to the directrix. Since FP = e · PM , by the definitionof the hyperbola, it follows that  FP2

    = e2

    ·PM 2; we may rewrite this equation

    in terms of coordinates as

    ( x  − ae)2 +  y 2 =  e2

     x  − ae

    2=  (ex  − a)2.

    Multiplying out the brackets we get

     x 2 − 2aex + a2e2 +  y 2 =  e2 x 2 − 2aex + a2,

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    32/602

    Conic Sections and Conics   17

    which simplifies to

     x 2

    e2 − 1

    −  y 2 = a 2

    e2 − 1

    ,

    or

     x 2

    a2 −   y2

    a2

    e2 − 1 = 1.Substituting b for  a

    √ e2 − 1, so that b2 =  a 2 e2 − 1, we obtain the standard

    form of the equation of the hyperbola

     x 2

    a2 −   y

    2

    b2 = 1.

    Notice that this equation is symmetrical in  x  and symmetrical in   y, so that  y

     x F F’   –a   afocus

    directrix

     x  = –a/edirectrix

     x  = a/e

    focus

    major axis

    minor axis

    b

    b

    the hyperbola also has a second focus   F (−ae, 0) and a second directrix   d with equation x  = −a/e.

    The hyperbola intersects the  x -axis at the points  (±a, 0). We call the seg-ment joining the points   (±a, 0)   the   major axis   or   transverse axis  of thehyperbola, and the segment joining the points  (0, ±b) the minor axis or conju-gate axis of the hyperbola (notice that this is NOT a chord of the hyperbola).The origin is the centre of this hyperbola.

    Notice also that each point with coordinates  (a sec t , b tan t ), where t  is notan odd multiple of  π/2, lies on the hyperbola, since

    In general,

    sec2 θ  = 1 + tan2 θ .

    a2 sec2 t 

    a2   − b2 tan2 t 

    b2   = 1.Then, just as for the parabola, we can check that

    Points for whicht  ∈  (−π/2, π/2)

    lie on the right branch of the hyperbola, and pointsfor which

    t  ∈  (π /2, 3π/2)lie on the left branch of the hyperbola.

     x  =  a sec t ,   y = b tan t    (t  ∈  (−π/2, π/2) ∪ (π/2, 3π/2))

    gives a parametric representation of the hyperbola.Two other features of the shape of the hyperbola stand out. Firstly, the

    hyperbola consists of two separate curves or  branches.Secondly, the lines with equations

     x 2

    a2 −   y

    2

    b2 = 0, or   y = ±b

    a x ,

    divide the plane into two pairs of opposite sectors; the branches of the hyper-

     y

    b

     –b

     x  = –a/e  x  = a/e y = – (b/a) x 

    F ′   –a   a F   x 

     y = (b/a) x 

    bola lie in one pair. As  x  → ±∞ the branches of the hyperbola get closer andcloser to these two lines. We call the lines  y = ±(b/a) x  the  asymptotes of thehyperbola.

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    33/602

    18   1: Conics

    We summarize the above facts as follows.

    Hyperbola in Standard Form   A hyperbola in standard form has equation   Another parametricrepresentation of thishyperbola is

     x  = a 1 + t 2

    1 − t 2 , y = b 2t 1 − t 2

    t  ∈ R− {±1}.

     x 2

    a2 −   y

    2

    b2 =  1, where b2 = a 2

    e2 − 1

    ,   a  > 0, e  >  1.

    It has foci  (±ae, 0) and directrices  x  = ±a/e; and it can be described bythe parametric equations

     x  =  a sec t ,   y = b tan t    (t  ∈  (−π/2, π/2) ∪ (π/2, 3π/2)).

    Problem 9   Let   P   be a point   (sec t ,   1√ 2 tan t ), where   (t    ∈   In this problem you will

    find the identity

    sec2 θ  = 1 + tan2 θ useful.

    (−π/2, π/2) ∪ (π /2, 3π/2)), on the hyperbola   E  with equation   x 2 −2 y2 = 1.(a) Determine the foci F  and  F  of  E .(b) Determine the slopes of   FP   and   F  P, when these lines are not

    parallel to the  y -axis.(c) Determine the point   P  in the first quadrant on   E  for which   FP is

    perpendicular to  F  P.

     y

     y = x 

     y= –x 

    b

     –aF ′   F   x a

     –b

    Rectangular Hyperbola ( e =√ 

    2) 

    When the eccentricity e  of a hyperbola takes the value √ 

    2, then  e2 =  2 andb =   a. Then the asymptotes of the hyperbola have equations   y  = ± x , sothat in particular they are at right angles. A hyperbola whose asymptotes are atright angles is called a rectangular hyperbola.

    Then, if we use the asymptotes as new x - and y -axes (instead of the original   We omit the details. x - and  y -axes), it turns out that the equation of the hyperbola can be written inthe form xy = c2, for some positive number  c.

     y

     x 

     xy= c2The rectangular hyperbola with equation xy = c2 has the origin as its centre,

    and the x - and y -axes as its asymptotes. Also, each point on it can be uniquelyrepresented by the parametric representation

     x  =  ct ,   y =   ct 

    where t  = 0.

    We shall use rectangular hyperbolas later on.   Section 2.5

    Polar Equation of a Conic 

    For many applications it is useful to describe the equation of a non-degenerateP

    O   Q  N 

    directrix d 

     M 

    q

    conic in terms of  polar coordinates r  and θ . A point   P( x , y) in the plane haspolar coordinates (r , θ ) if  r  is the distance  OP (where  O  is the origin) and θ  isthe anticlockwise angle between OP and the positive direction of the  x -axis.

    Take the origin   O  to be the focus of the conic,  d  the directrix,   M  the footof the perpendicular from a point   P  on the conic to   d ,   N   the foot of theperpendicular from O  to  d , and Q  the foot of the perpendicular from  P  to  ON .

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    34/602

    Conic Sections and Conics   19

    Then by the definition of the conic, we have  OP =  e · PM . We can rewritethis as

    r  =  e(ON − OQ)=  e · ON − er  cos θ ,

    orr (1 + e cos θ ) = e · ON 

    = l , a constant.It follows that the equation of the conic can be expressed in the form

    r  =   l1 + e cos θ .

    The polar form of the equation of a conic is often used in problems inplanet

    Sun Sun

    comet

    Dynamics: for example, in determining the motion of a planet or of a cometround the Sun.

    1.1.4 Focal Distance Properties of Ellipse and Hyperbola

    We now prove two simple but surprising results. We deal with the ellipse first.

    Theorem 5 Sum of Focal Distances of Ellipse

    Let  E  be an ellipse with major axis  (−a, a) and foci  F  and  F . Then, if   Pis a point on the ellipse, FP + PF  = 2a. In particular, FP + PF  is constantfor all points  P on the ellipse.

     y

     x 

    directrix d directrix d ′

    P

    F ′   F 

    Proof    Let d  and  d  be the directrices of the ellipse that correspond to the fociF  and  F , respectively. Then, since

    PF  = e × (distance from  P to  d )and

    PF  = e × (distance from  P to  d ),

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    35/602

    20   1: Conics

    it follows that

    PF + PF  = e × (distance between d  and  d ).= 2a,

    which is a constant.  

    The result of Theorem 5 can be used to draw an ellipse, using a piece of 

    pencil

    ellipse

    fixed endsstring fixed at both ends. A pencil is used to pull the string taut; then, as wemove the pencil round, the shape that it traces out is an ellipse whose foci arethe two ends of the string.

    Notice that, if we are given any three points   F ,   F   and   P  (not on the linesegment   F  F ) in the plane, then there is only one ellipse through   P  with   F    For, the location of the

    foci and the length of themajor axis specify anellipse uniquely.

    and   F  as its foci. Its centre is the midpoint,   O, of the segment   F  F , its axesare the line along   F  F  and the line through   O  perpendicular to   F  F , and itsmajor axis has length PF + PF .

    Also, if we are given any two points   F   and   F   in the plane, the locus of    As a result, some bookstake the ‘Sum of FocalDistances Property’ as thedefinition of the ellipse.

    points   P (not on the line segment   F  F ) in the plane for which  PF + PF  is aconstant is necessarily an ellipse. Thus the converse of Theorem 5 holds.

    There is an analogous result for the hyperbola.

    Theorem 6 Difference of Focal Distances of Hyperbola

    Let  H  be a hyperbola with major axis (−a, a) and foci  F  and   F . Then, if P is a point on the branch of the hyperbola that is closer to  F ,

    PF  − PF  = 2a;and, if  P is a point on the branch of the hyperbola closer to  F ,

    PF  − PF  = −2a.In particular, |PF  − PF | is constant for all points  P on the hyperbola.

     y

    P

     x F F ′

    directrix d ′ directrix d 

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    36/602

    Conic Sections and Conics   21

    Proof    We shall prove only the first formula; the proof of the second is similar.Let d  and  d  be the directrices of the hyperbola that correspond to the foci

    F  and  F  respectively, and let  P be a point on the branch of the hyperbola thatis closer to  F . Then, since

    PF  = e × (distance from  P to  d )

    and

    PF  = e × (distance from  P to  d ),

    it follows that

     X 

    P

    F F ′

    pencil

    PF  − PF  = e × (distance between d  and  d )= 2a,

    which is a constant.  

    The result of Theorem 6 can be used to draw a hyperbola, this time using apiece of string and a stick. Choose two points  F  and  F  on the x -axis, equidis-tant from and on opposite sides of the origin. Hinge one end of a movablestick   F  X  at the focus   F ; attach one end of a string of length   (where    isless than the length of  F  X ) to the end  X  of the stick and the other end of thestring to  F , and keep the string taut by holding a pencil tight against the stick,as shown.

    Then, as we move the pencil along the stick, the shape that it traces out ispart of one branch of a hyperbola with foci  F  and  F . For,

    PF 

     −PF 

     = XF 

     −( XP

    +PF )

    =  XF  − = a constant independent of  P.

    We obtain the other branch of the hyperbola by interchanging the roles of   F and  F  in the construction.

    Notice that, if we are given any three points   F ,   F  and   P  (not on the linethrough F  F  or its perpendicular bisector) in the plane, then there is only one   For, the location of the

    foci and the length of themajor axis specify a

    hyperbola uniquely.

    hyperbola through   P  with   F  and   F  as its foci. Its centre is the midpoint,   O,of the segment   F  F , its axes are the line along   F  F  and the line through   O

    perpendicular to  F  F , and its major axis has length |PF  − PF |.Also, if we are given any two points   F   and   F   in the plane, the locus of 

    points   P  (not on the line segment   F  F ) in the plane for which PF  − PF  is anon-zero constant is necessarily one branch of a hyperbola. Thus the converse   As a result, some books

    take the ‘Difference of Focal Distances Property’as the definition of thehyperbola.

    of Theorem 6 holds, in the following sense: Given any three points  F ,  F  andP   (where   P  must lie strictly between   F  and   F   if it lies on the line throughF  F ) in the plane for which PF  − PF  = 0, the locus of points  Q in the planefor which QF  − QF  = ±|PF  − PF | is a hyperbola.

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    37/602

    22   1: Conics

    1.1.5 Dandelin Spheres

    We now give a beautiful proof due to Dandelin of the fact that a slant plane   Germinal Pierre Dandelin(1794–1847) was aFrench–Belgian Professorof Mechanics at LiègeUniversity; he made his

    discovery in 1822.

    π  that cuts one portion of a right circular cone in a ‘complete’ curve   E  is anellipse, just as it appears to be!

    To do this, first fit a sphere inside the cone so that touches the plane  π   (at

    a point   F ) and the cone (in a circle  C  with centre   O), as shown in the figurebelow. The circle  C  lies in a horizontal plane which intersects  π  in a line  d .Take an arbitrary point   P on the curve   E , and extend the line from the vertexV  of the cone through  P to meet C  at the point  L , and let  D be the point on  d such that PD is perpendicular to d .

    The line PD lies in a vertical plane which intersects the vertical plane  VLOin the line PM , so that PMD and PML are both right-angled triangles.

    Denote by  α  the angle between the slant plane  π  and the horizontal planethrough C , and by β  the angle ∠PLM  (the base angle of the cone).   We use the symbol ∠ to

    indicate an angle.

     E 

    F P

     D

    O

    C  M 

     L

    a

    b

    p

    From the right-angled triangles PMD and PML, we see that

    PL =   PM sin β

    and   PD =   PM sin α

    ,

    so thatPL

    PD = sin α

    sin β .

    Now   PF  =   PL  since they are both tangents from a given point to a givensphere; it follows that

    PF 

    PD=  sin α

    sin β. (11)

    Now 0  < α < β <   π2 , since the plane  π  is less steep than the base angle of the cone. So if we let  e = sin α/ sin β, it follows that 0  <  e  <  1.

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    38/602

    Properties of Conics   23

    It then follows from equation (11) that for any point   P  on the curve   E ,   V d ′

    F ′

    p

    its distance   PF   from the fixed point   F   is   e   times its distance   PD   fromthe fixed line   d . Since 0 < e < 1, it follows from the focus–directrix defini-tion of an ellipse that the curve   E  must be an ellipse, with focus   F   anddirectrix d .

    If we then construct the other sphere that touches both  π  and the cone, asimilar argument shows that the point of contact of the sphere with  π  is theother focus   F   of the ellipse; and the other directrix of the ellipse is the lineof intersection of  π  with the horizontal plane through the circle in which thesphere touches the cone.

    A similar construction involving spheres proves that in the cases 4 and6 illustrated in the sketch in Subsection 1.1.1 the curve of intersection is a   We ask you to look at the

    two cases 4 and 6 inExercise 7 of Subsection 1.5.

    parabola and a hyperbola, respectively.This completes the proof of our claim in Subsection 1.1.1 that the curves of 

    intersection of certain planes with a double cone are an ellipse, a parabola or ahyperbola. We shall investigate the converse in Theorem 4 of Subsection 4.1.4.

    1.2 Properties of Conics

    1.2.1 Tangents

    In the previous section you met the parametric equations of the parabola,   Subsection 1.1.3ellipse and hyperbola in standard form.

    We now tackle a rather natural question: given parametric equations   x  = curve

    tangent

    ( x (t ), y (t ))

     x (t ),  y =  y(t ) describing a curve, what is the slope of the tangent to the curveat the point with parameter  t ? This information will enable us to determine theequation of the tangent to the curve at that point.

    ( x (t ), y (t ))

    ( x (t  + h), y (t  + h))

     y(t  + h)

    – y (t )

     y (t  + h) – x (t )

    Theorem 1   The slope of the tangent to a curve in  R2 with parametricequations x  =  x (t ),  y =  y(t ) at the point with parameter t  is

     y(t ) x (t )

    ,

    provided that x (t ) = 0.

    Proof    The points on the curve with parameters t  and  t  + h have coordinates( x (t ), y(t )) and ( x (t 

     +h), y(t 

     +h)), respectively. Then, if  h

     = 0, the slope of 

    the chord joining these two points is

     y(t  + h) − y (t ) x (t  + h) −  x (t ) ,

    which we can write in the form

    ( y(t  + h) −  y (t ))/h( x (t  + h) −  x (t ))/h .

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    39/602

    24   1: Conics

    We then take the limit of this ratio as  h →  0. The slope of the chord tends tothe slope of the tangent, namely  y (t )/ x (t ).  

    Example 1

    (a) Determine the equation of the tangent at the point with parameter  t  to the

     –a

     –b

    b

     y

     x a

    ellipse with parametric equations

     x  =  a   cos   t ,   y = b  sin   t ,

    where t  ∈  (−π , π], t  = 0,  π .(b) Hence determine the equation of the tangent to the ellipse with parametric

    equations  x  =  3cos   t ,   y =   sin   t  at the point with parameter   t  =   π/4.Deduce the coordinates of the point of intersection of this tangent with the x -axis.

    Solution 

    (a) Now,   y(t ) =   b   cos   t  and  x (t ) = −a   sin   t   for  t  ∈   (−π , π]; it followsthat, for  t  =  0 or  π , the slope of the tangent at the point with parametert  is

     y(t ) x (t )

    =   b  cos   t −a   sin   t .

    Hence the equation of the tangent at the point  (a   cos   t ,   b   sin   t ),  t  =  0,π , is

     y − b   sin   t  = −b  cos   t 

    a sin   t  ( x  −   a cos   t ).Multiplying both sides and rearranging terms, we get

     xb cos   t  + ya  sin   t  =  ab  cos2 t  + ab  sin2 t  =  ab,

    and dividing both sides by ab gives the equation

    We shall use this equationin Subsection 1.2.2.

     x 

    acos   t  +   y

    bsin   t  = 1. (1)

    The point on the ellipse where  t  =  0 is (a, 0), at which the tangent hasequation x 

     = a . Similarly, the point on the ellipse where  t 

     = π  is (

    −a,0),

    at which the tangent has equation   x  = −a. It follows that equation (1)covers these cases also.

    (b) Here the curve is the ellipse in part (a) in the particular case that  a =  3,b =  1. When t  =   π/4, it follows from equation (1) that the equation of the tangent at the point with parameter t  =  π /4 is

     x 3 .

      1√ 2 + y ·   1√ 

    2 = 1,

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    40/602

    Properties of Conics   25

    or13 x  + y =

    √ 2.

    Hence, at the point  T  where the tangent crosses the  x -axis,   y =  0 and so x  = 3

    √ 2. Thus, T  is the point

    3√ 

    2, 0

    .

    Problem 1   Determine the slope of the tangent to the curve inR2 with y

     x 

    cardioid

     x  = 2 cos t  + cos 2t  + 1

     y = 2 sin t  + sin 2t 

    parametric equations

     x  = 2 cos t  + cos2t  + 1,   y = 2 sin t  + sin2t 

    at the point with parameter   t,   where   t  is not a multiple of   π . Hencedetermine the equation of the tangent to this curve at the points withparameters t  =  π/3 and t  =  π/2.

    Problem 2

    (a) Determine the equation of the tangent at a point  P with parametert  on the rectangular hyperbola with parametric equations   x  =   t , y = 1/t .

    (b) Hence determine the equations of the two tangents to the rectangular

     y

     xy = 1

     x (1, –1)

    P(t ,1–t )

    hyperbola from the point (1, −1).

    We can modify the result of Example l(a) to find the equation of the tangent

    at the point  ( x 1, y1) on the ellipse with equation   x 2

    a2 +   y2

    b2 =  1. We take   x  =

    a cos t ,  y

     = b sin t  as parametric equations for the ellipse, and let x 1

     = a cos t 1

    and  y1 = b sin t 1. Then it follows from equation (1) above that the equation of the tangent is

     x 

    acos t 1 +

      y

    bsin t 1 = 1,

    which we can rewrite in the form   xx 1a2

     +   yy1b2

      = 1.We can determine the equations of tangents to the hyperbola and the

    parabola in a similar way; the results are given in the following theorem.

    Theorem 2   The equation of the tangent at the point ( x 1, y1) to a conic instandard form is as follows.

    Conic Tangent  

    Ellipse  x 2

    a2 +   y

    2

    b2 = 1   xx 1

    a2 +  yy1

    b2 = 1

    Hyperbola  x 2

    a2 −   y

    2

    b2 = 1   xx 1

    a2 −  yy1

    b2 = 1

    Parabola y 2 = 4ax yy1 = 2a( x  + x 1)

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    41/602

    26   1: Conics

    Problem 3   Prove that the equation of the tangent at the point ( x 1, y1)to the rectangular hyperbola  xy = 1 is   12 ( xy1 + x 1 y) = 1.

    Problem 4   For each of the following conics, determine the equation(a, b)

    ( x 2 , y

    2)

    ( x 1 , y1)

    of the tangent to the conic at the indicated point.

    (a) The unit circle  x 2 + y 2 = 1 at− 12 ,  12

    √ 3

    .

    (b) The hyperbola x y = 1 at−4, − 14

    .

    (c) The parabola  y 2 =  x  at  (1, −1).

    We can deduce a useful fact from the equation  xx 1 + yy1 =  1 for the tangentat the point ( x 1, y1) to the unit circle x 2 + y 2 = 1. Let (a, b) be some point onthis tangent, so that

    ax 1

     +by1

     = 1. (2)

    Next, let the other tangent to the unit circle through the point  (a, b) touch thecircle at the point ( x 2, y2); it follows that   This is because (a, b) lies

    on the tangent at ( x 2, y2),whose equation is xx 2 + yy2 = 1.

    ax 2 + by2 = 1. (3)

    From equations (2) and (3) we deduce that the points  ( x 1, y1) and  ( x 2, y2)both satisfy the equation  ax  + by =  1. Since this is the equation of a line, itmust be the equation of the line through the points  ( x 1, y1) and  ( x 2, y2). Forhistorical reasons, this line is called the polar of   (a, b) with respect to the unit    We shall meet polars of 

    other conics inSubsection 4.2.1.

    circle.

    Theorem 3   Let (a, b) be a point outside the unit circle, and let the tangents

    polar of 

    (a, b)

    (a, b)

    P1

    P2

    curve

    normaltangent

    P

    to the circle from (a, b) touch the circle at   P1 and  P2. Then the equation of the line through  P1 and  P2 is

    ax + by = 1.

    For example, the polar of (2, 0) with respect to the unit circle is the line2 x  = 1.

    Problem 5   Determine the equation of the polar of the point (2, 3) withrespect to the unit circle.

    In the next example we meet the idea of the normal to a curve.

    Definition   The normal to a curve  C  at a point P on  C  is the line throughP that is perpendicular to the tangent to C  at P.

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    42/602

    Properties of Conics   27

    Example 2

    (a) Determine the equation of the tangent at the point with parameter t  to the y

     x 

    (at 2 , 2at )

    parabola with parametric equations

     x  =  at 2,   y = 2at    (t  ∈ R).(b) Hence determine the equations of the tangent and the normal to the

    parabola with parametric equations   x  =   2t 2,   y =   4t  at the point withparameter t  = 3.

    Solution 

    (a) Since   y(t ) = 2a and  x (t ) =  2at , it follows that, for  t  = 0, the slope of the tangent at this point is

     y(t ) x (t )

    =   2a2at 

    =  1t 

    .

    Hence the equation of the tangent at the point  (at 2, 2at ), t  = 0, is y − 2at  = 1

    t ( x  − at 2),

    which can be rearranged in the form

    We shall use this equationin Subsection 1.2.2.

    ty = x  + at 2. (4)

    The point on the parabola at which  t  = 0 is (0, 0); there the tangent to theparabola is the   y-axis, with equation  x  =  0. It follows that equation (4)covers this case also.

    (b) Here the curve is the parabola in part (a) in the particular case that  a = 2.When t  =  3, it follows from equation (4) that the equation of the tangentis 3 y = x  + 2 · 32, or 3 y =  x  + 18.

    To find the equation of the normal, we must find its slope and the   Recall that lines of (non-zero) slope m1 andm2 are perpendicular if and only if  m1 · m2 = −1.

    coordinates of the point on the parabola at which  t  = 3.When t  =  3, it follows from the equation of the tangent that the slope

    of the tangent is   13 . Since the tangent and normal are perpendicular to eachother, it follows that the slope of the normal must be −3. Also, when t  = 3,we have that  x  =  2 · 32 =  18 and   y =  4 · 3 =  12; so the correspondingpoint on the parabola has coordinates (18, 12).

    It follows that the equation of the normal to the parabola at the point

    (18, 12) is y − 12 = −3( x  − 18)

    = −3 x  + 54,or

     y = −3 x  + 66.

  • 8/20/2019 Geometry David a Brannan Matthew f Esplen Jeremy j Gray

    43/602

    28   1: Conics

    Problem 6   The normal to the parabola with parametric equations y

     x 

    Q(T 2 , 2T )

    P(t 2 , 2t )

     x  =  t 2,  y = 2t  (t  ∈ R) at the point P with parameter t , t  = 0, meets theparabola at a second point Q with parameter T .

    (a) Prove that the slope of the normal to the parabola at P is −t .(b) Find the equation of the normal to the parabola at  P.(c) By substituting the coordinates of   Q  into your equation from part

    (b), prove that  T  = −2t  − t .

    Problem 7   This question concerns the parabola with parametricequations x  =  at 2,  y = 2at  (t  ∈ R).

    directrix

    F (a, 0)

    P2

    P1

     yQ

     x 

     –a

    (a) Determine the equation of the chord joining the points P1 and  P2 onthe parabola with parameters  t 1 and  t 2, respectively, where t 1 and  t 2are unequal and non-zero.

    Now assume that the chord   P1 P2 passes through the focus  (a, 0) of theparabola.

    (b) Prove that t 1t 2 = −1.(c) Use the result of Example 2(a) to write down the equations of the

    tang